Who is in control?
Look for parent-managed permissions. Wondoki offers individual limits, schedules, device revocation and optional-tool switches.
SAFETY, EXPLAINED
Curiosity needs
careful boundaries.
Wondoki combines parent choices with checks before and after AI answers. These measures reduce risk. They do not guarantee that every answer is correct or appropriate.
BEFORE AN ANSWER APPEARS
Several checks. A clear limit.
- Read the current family rules. The server checks the profile, age, allowed time and feature permissions.
- Check the question and context. Safety checks can redact information, redirect a request or withhold an answer. They can also misunderstand an ordinary question.
- Generate for the child’s age. Instructions guide vocabulary, detail and homework help. They do not guarantee a perfect result.
- Check before delivery. The app buffers answers until output checks finish. If a required check fails or is unavailable, the answer is withheld.
PARENT CHOICE
Six levels. Some boundaries stay fixed.
Safety levels adjust how cautiously the assistant approaches sensitive subjects. They do not enable harmful instructions or remove core protections. A factual, non-graphic body-education question is different from sexualized content; a science question is different from instructions to hurt someone. Context, wording and settings matter.
Review examples in each child’s settings. Report answers that miss the mark.
CHOOSING AN AI TOOL
What should a parent check?
General-purpose chatbots and open search can expose children to adult topics, unreliable claims, persuasive conversation or unsuitable links. Features vary by product and plan. Check each service’s current age rules and privacy settings.
Wondoki is built around family controls. It still shares important AI risks: mistakes, bias, missed safety signals and overreliance. We do not claim that every other service is dangerous or that Wondoki eliminates those risks.
Where can a question go?
Wondoki search is limited to filtered Wikipedia results. External pages have their own content and can change after a link is shown.
What happens to the conversation?
Look for clear provider, storage and deletion information. Wondoki offers retention choices and exports; parents can read retained chats.
The boundaries worth knowing
Can a safety alert miss something?
Wondoki automatically checks questions and AI answers during each conversation. These checks can miss risks or flag harmless content. Alerts are not continuously monitored by a human, and Wondoki is not an emergency-response service.
Can AI become a substitute for real people?
It should not. Wondoki identifies itself as an AI learning assistant. Encourage offline activities and trusted adults. A warm tone is interface design, not a human relationship.
Are voice and pictures private?
They are optional and processed by an external provider when enabled. Photos are resized locally. Wondoki does not retain originals in its database, but prepared text can become part of the chat. Keep people, identifying details and private documents out of uploads.
What if an answer feels wrong?
Stop and check it together. Change settings or disconnect the device if necessary. Email info@tallinio.com with a short description; avoid forwarding a full private transcript.
Is Wondoki GDPR- or COPPA-certified?
We do not claim certification or completed legal compliance. Verified parental consent, processor arrangements, retention and specialist evaluations remain launch requirements. A privacy notice does not complete them.
Something you want to understand?
Contact Tallinio LTD at info@tallinio.com. For an immediate risk to someone’s safety, use appropriate local emergency or safeguarding services.
Privacy, in detail